some fix

app.py

@@ -101,11 +101,11 @@ def inference(img):
         generation_kwargs = dict(
             **inputs,
             streamer=streamer,
-            max_new_tokens=128, # Increased slightly
+            max_new_tokens=128,
             min_new_tokens=2,
-            do_sample=False,
-            use_cache=False,
-            temperature=1.5, # Adjusted slightly for stability
+            do_sample=True,
+            use_cache=True,
+            temperature=1.5,
             min_p=0.1,
         )
 
@@ -114,8 +114,8 @@ def inference(img):
         thread.start()
 
         generated_text = ""
-        status_indicator = "   **...generating** ⏳"
-        yield " ** The text will start showing shortly... ⏳**"
+        status_indicator = "   ...generating ⏳"
+        yield "The text will start showing shortly... ⏳"
         for new_text in streamer:
             if generated_text == "":
                 # Todo: remove the removal of first character when the model is fixed